According to Vertical Systems Group, customer installations of carrier-based fame relay services in 1991 consisted of 120 ports worldwide. By 2001, the number of ports had increased to approximately 1.78 million. (http://www.verticalsystems.com/press15.html, Vertical Systems Group, Inc., Dedham, Mass.). Providers often route frame relay services over asynchronous transfer mode (ATM) backbones. According to the ATM Form (www.atmforum.com), approximately eighty percent of the world's telecommunications service providers use ATM in the core of their networks. (Beginners' Overview of Asynchronous Transfer Mode (ATM), http://www.atmforum.com/pages/aboutatmtech/beginnersguide.html). As the number of customer and network installations has increased, so has the difficulty of effectively and efficiently monitoring and managing the networks supporting the providers' customers.
Service providers and equipment vendors have attempted to address this increasing difficulty in a variety of ways. One approach taken by service providers has been to over-engineer communications networks in an attempt to alleviate problems resulting from the over-utilization of the frame relay and ATM network resources and corresponding decreases in the quality of service. Although over-engineering the network may help the provider ensure that their customer receives an adequate level of service, this approach is both expensive and an inefficient allocation of scarce resources, including both capital and labor.
Another approach that attempts to ensure at least a minimum customer service level is monitoring the frame relay and ATM networks at the end-points of a communication to determine if the network is performing adequately. Conventional approaches to monitoring the network end-points require additional hardware, additional software, or both. Hardware approaches utilize probes, which are installed between the network and the customer premises equipment (CPE) or between the network and a router or switch at the provider's facility. A probe may be a stand-alone network element or may be a hardware and/or software package integrated into a switch. For example, a Remote Monitoring (RMON) probe is a device or software agent that collects and compiles data about a network segment and transfers the collected information to a management workstation on request. RMON probes are generally utilized in TCP/IP and Ethernet environments.
Conventionally, probes communicate the network monitoring data to a network management workstation. The workstation may communicate a request to the probe, or the probe may periodically send data to the workstation. The workstation does not have to be on the same network as the probe and can manage the probe by in-band or out-of-band connections.
In addition to using separate network elements, such as probes, a provider may utilize a network element that incorporates software directly on the element. The software causes the network element to capture monitoring information and to send it to a network management workstation periodically or on request in a manner similar to that implemented by a hardware-based probe.
The installation of conventional probes, both hardware and software-based, on the network addresses the need to monitor and manage the network. However, this approach entails several disadvantages. Utilizing this approach requires installing either an additional physical device or installing additional monitoring software on an existing device. Additionally, not all network devices are compatible with monitoring software, and therefore, this approach may require the purchase of new network devices, such as switches, if a software approach is utilized. Also, this approach entails substantial expense, including the purchase cost of the probes, both hardware and software-based, the installation time and cost of the probe hardware and software at the various customer locations or other endpoints of a communication, and the ongoing support cost, including software upgrade costs.
Another approach to monitoring the network is to analyze data as it arrives at the customer premises directly. One method for a network administrator, analyst, or automated system (hereinafter “analyst”) to perform localized analyses is by setting up a mirror for a port, either the input or output port, on the customer premises equipment to another port on the same network element. Localized monitoring eliminates the additional cost required to install a probe and remote monitoring software. However, this approach requires the presence of an analyst to set up the mirror and to monitor the network. In addition, monitoring can only take place where the customer premises or other communications end-point equipment is physically located. And therefore, the analyst is unable to monitor both ends of the communication simultaneously. Localized monitoring is inefficient and expensive, particularly in terms of the analyst's valuable time.
None of the conventional approaches provides a means of efficiently and effectively monitoring and managing network communications end-points.